INDEX

Explanations

phrases related to transformations or improvements

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 constitu

-0.67

 Bastard

-0.63

selves

-0.61

aples

-0.61

 surv

-0.58

 withd

-0.58

 davidjl

-0.58

 sworn

-0.57

eger

-0.57

 destro

-0.57

POSITIVE LOGITS

ings

1.74

able

1.70

ables

1.56

away

1.47

ability

1.47

aways

1.42

outs

1.36

downs

1.34

ers

1.32

backs

1.28

Activations Density 3.247%

phrases related to transformations or improvements

No Comments

No Known Activations